Fearless Steps Challenge:
Phase II

Conversational Analysis

Identifying salient events in a continuous audio stream is important. Its significance is even more for time-critical events such as the Apollo Missions. Identification of ‘hotspots’ in over 150,000 hours can help both STEM and non-STEM researchers in analysing high value content. We define the following critical events as ‘hotspots’

• Alarms
• Loss of Signal
• Acquisition of Signal
• mission critical status checks
• mission landings (earth-reentry and lunar)
• mission liftoffs (earth and lunar)
• Team switch
• Checkpoints

These hotspots are defined a conversational context. All mission control speech segments with the above mentioned events as primary topics of discussion between personnel are considered for this task.

Instructions

Performance Metrics

Both tracks in the Conversational Analysis Task will be evaluated for Accuracy of the Top-3 system predictions for a given input file.

where, S = {k ∈ [1, M] : Nref (k) ⊆ Nsys(k)} and M is the total number of input segments.

For additional details about labeling, please consult the documentation.